Methodology

Every threshold on this page applies to every figure the site publishes, without exception. If a figure appears on the site, it passed these rules; if it does not appear, one of them stopped it.

Where facts come from

Sources are ranked, and a fact can never claim more certainty than its source supports. The hierarchy, strongest first:

1. The operator’s official sweepstakes rules.
2. Official redemption or payments documentation.
3. Official help-centre material, or a directly observed response from the operator’s own servers.
4. A direct response from operator support.
5. Redacted user evidence that we can check.
6. An unverified user report.
7. Secondary reporting — anything written about the operator by a third party.

The ceiling is absolute. A secondary write-up — the weakest kind of source we record — cannot produce an “official” label no matter how confidently it is written, and such a record is never published. This is the rule that stops a comparison site’s claim about a minimum from becoming, on our pages, the operator’s published minimum.

We record failed retrievals as carefully as successful ones. “We fetched the operator’s rules and they could not be read” and “we never looked” are very different statements, and both appear in the source list with the date and the exact path attempted.

Published against observed

A published figure is what the operator states in its own documentation. An observed figure is what a set of recorded redemptions actually did. These are never merged, never averaged together, and never displayed in a way that lets one be mistaken for the other. Published versus observed · First versus repeat · A maximum is not typical · Failed retrievals.

Where an operator publishes a maximum — “up to ten days” — we show it as a maximum. Converting a stated ceiling into a typical time would be inventing a claim the operator never made.

How observed timings are calculated

• Only observations that are published, current, actually completed, and carrying an elapsed duration are eligible. Pending redemptions have no duration yet, and counting them would pull every figure downward.
• Observations are never pooled across redemption methods or currencies. A gift card and a bank transfer are different products.
• First and repeat redemptions are reported separately. The first usually includes identity verification and the repeat usually does not, so a pooled figure describes an experience nobody actually has.
• We report the median with a range, never the mean. A single long outlier moves a mean far enough to misrepresent the typical case.
• The median and the quartiles use the nearest-rank method, so every figure shown is a duration some redemption actually took rather than an interpolated half-day no redemption produced. An even-sized sample therefore reports one of the two central observed durations, not their average.

Sample thresholds

Below 5 eligible observations we publish no central figure at all, and say so in place of the number. From 5 through 11 the figure is published and labelled a small sample. At 12 or more the small-sample caveat is removed — the sample is treated as confident — but the sample size itself is still displayed, on every figure, everywhere on the site.

5 is not a statistically strong sample. It is the point below which one record would drive the whole published number.

Outliers

A point more than 1.5 times the interquartile range beyond the middle half is flagged — and kept. It stays in the median, in the range and in the sample count.

We do not trim outliers. On a site about whether people get paid, the redemption that took ninety days is the single most informative record in the set, and quietly dropping it to tidy up a distribution would be the exact failure this site exists to avoid.

Freshness

Redemption terms change without announcement. A fact verified within 45 days is presented plainly. Past that it is labelled as needing re-checking, and past 120 days it is labelled out of date. The label is computed from the verification date, so a fact ages visibly whether or not anyone revisits it.

A verification date is the date a human checked the source. It is never the date the site was rebuilt — a rebuild re-verifies nothing.

What we deliberately do not do

• No trust score, no letter grade, no composite rating. Collapsing minimums, timings, verification requirements and evidence quality into one number destroys the only information a reader needs and hides how much of it we actually know.
• No red/green safety badge. A colour-coded verdict on thin evidence is a judgement dressed as a measurement.
• No site-wide average cashout time. It would blend methods, operators and sample sizes into a number that describes nothing.
• No prediction. Observations describe the redemptions in the sample. They do not forecast what a future redemption will do, and no wording on this site should suggest otherwise.
• No ranking by payout, by commission, or by anything else. We do not order operators by quality at all.

Corrections

Any figure here can be challenged against a citable source, by an operator or by a reader. Corrections are recorded with what was challenged, what the evidence was, and what changed. See corrections and the evidence policy.